SahilCarterr/Text-to-Python-Code-Generator

1import gradio as gr2from transformers import AutoModelForCausalLM, AutoTokenizer3from peft import PeftModel, PeftConfig4 5# Load the PEFT configuration, base model, and tokenizer6config = PeftConfig.from_pretrained("SahilCarterr/Llama-2-7B-Chat-PEFT")7base_model = AutoModelForCausalLM.from_pretrained("TheBloke/Llama-2-7b-Chat-GPTQ", device_map='auto')8model = PeftModel.from_pretrained(base_model, "SahilCarterr/Llama-2-7B-Chat-PEFT")9tokenizer = AutoTokenizer.from_pretrained("SahilCarterr/Llama-2-7B-Chat-PEFT")10 11def respond(message, history, system_message, max_tokens, temperature, top_p):12    messages = [{"role": "system", "content": system_message}]13 14    for val in history:15        if val[0]:16            messages.append({"role": "user", "content": val[0]})17        if val[1]:18            messages.append({"role": "assistant", "content": val[1]})19 20    messages.append({"role": "user", "content": message})21    22    # Encode the input23    inputs = tokenizer(message, return_tensors="pt").input_ids.to('cuda')24    # Generate the response using the model25    outputs = model.generate(inputs, max_new_tokens=max_tokens, do_sample=True, temperature=temperature, top_p=top_p)26    response = tokenizer.decode(outputs[0], skip_special_tokens=True)27 28    yield response29 30demo = gr.ChatInterface(31    respond,32    additional_inputs=[33        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),34        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),35        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),36        gr.Slider(37            minimum=0.1,38            maximum=1.0,39            value=0.95,40            step=0.05,41            label="Top-p (nucleus sampling)",42        ),43    ],44)45 46if __name__ == "__main__":47    demo.launch()